Notes about the upload

Created a Wikidata item to source the edits: Q68644488.

Excluded 11 rows containing variations of 'test'.

Name of community reconciled against type "human settlement" with country column as P17

Reconciliation against name – 23 items matched. Looked messy – e.g. several rows with different coordinates were matched to KNUST Library Q19462815. Decided to not process matched items.

Left with 821 items to be created. 660 were created in Quick Statements, 161 generated errors and were not created.

The upload revealed a number of rows in the source data have incorrect coordinates.

Re: incorrect coordinates, it doesn't look like the numbers are in the wrong order, but they're completely wrong.

E.g. https://www.wikidata.org/wiki/Q68651778 has in the source data
y (latitude) 62
x (longitude) 6

While another library in the same town, showing correctly on the map, has
y (latitude) 5
x (longitude) -0.1

So it might make more sense to completely remove the coordinates statements that look wrong rather than try to figure out how exactly they are wrong.

Manually removed the coordinates from 15 items where they looked obviously wrong on the map:

Q68653096
Q68651910
Q68652604
Q68651882
Q68653101
Q68651603
Q68653106
Q68653087
Q68651930
Q68652603
Q68651821
Q68653302
Q68651696
Q68652831
Q68651778

Could you add a link to the sparql query for "all libraries in Africa with this dataset colour coded differently" which you shared in the hangout?

@Lokal_Profil it is already there :) On a map, all libraries in Africa colored according to "stated in" reference for coordinates (highlighting those with coords imported from this dataset)

In T233902#5529423, @Alicia_Fagerving_WMSE wrote:

@Lokal_Profil it is already there :) On a map, all libraries in Africa colored according to "stated in" reference for coordinates (highlighting those with coords imported from this dataset)

Sorry about that. I blame Friday blindness =)